The Cambridge textbook of bioethics

74B. M. Dickens, J. M. Boyle Jr., and L. Ganzini

suicide remain offences in most jurisdictions. However, several jurisdictions such as the Netherlands and Belgium are coming to recognize individuals’ capacity for rational choice of suicide, and the right of physicians to give assistance.

A concern regarding approaching euthanasia and medically assisted suicide through criminal law is that enforcement may be ineffective. Physicians may be justified in increasing medications for pain control, as patients’ relief from pain at given dosage levels decreases, until a toxic level is predictably reached and is the precipitating cause of death. Patients’ deaths result, however, not from their treatment but from their pathologies, which justified and even compelled the pain relief treatment (R v. Adams, 1957; Williams, 2001). Physicians who withhold indicated measures of pain relief for fear of personal accountability for their patients’ deaths are in a conflict of interest. However, prosecutors may find it impossible to show beyond reasonable doubt that physicians’ primary intentions are not pain relief but ‘‘mercy killing.’’

Similarly, medications may properly be prescribed for patients’ periodic self-administration, which they may hoard and then consume at the same time in order to commit suicide. Physicians may recognize this as a risk, but it may be impossible to show beyond reasonable doubt that they intended this consequence or were negligent. Warning patients of dangers to their lives of over medication may send an ambiguous message.

Empirical studies

A study in 1995 in Canada (Singer et al., 1995) showed that more than 75% of the general public supported voluntary euthanasia and assisted suicide in the case of patients who were unlikely to recover from their illness. Roughly equal numbers, however, opposed these practices for patients with reversible conditions (78% opposed), elderly disabled people who feel they are a burden to others (75% opposed), and elderly people with minor physical ailments (83% opposed) (Genuis et al., 1994). Results of one survey indicated that 24% of

Canadian physicians would be willing to practice euthanasia and 23% would be willing to assist in a suicide if these acts were legal (Wysong, 1996). These findings are similar to the results of surveys conducted in the UK (Ward and Tate, 1994) and in Australia’s Northern Territory (Anon., 1996). Surveys of physicians in the Australian state of Victoria (Kuhse and Singer, 1988), as well as surveys in Oregon (Lee et al., 1996), Washington (Shapiro et al., 1994), and Michigan (Bachman et al., 1996) indicated that a majority of physicians in these jurisdictions supported euthanasia and assisted suicide in principle and favored their decriminalization. Physicians in certain specialties, such as palliative care, appear to be less willing to participate in euthanasia and assisted suicide than physicians in other specialties.

Approximately 3% of all deaths in the Netherlands result from euthanasia or assisted suicide (van der Maas et al., 1996). Most of these patients have cancer, though one in five patients with amyotrophic lateral sclerosis die of euthanasia or assisted suicide (Veldink et al., 2002). Physicians report that Dutch patients pursue euthanasia because of loss of dignity, ‘‘unworthy dying,’’ and dependence on others. Pain was mentioned as a reason for pursuing hastened death by almost half of patients, but in only 5% was it the sole reason (van der Maas et al., 1996).

In a national US sample of almost 2000 physicians, one in six reported having received a request from a patient for assistance with suicide; 11% had received a request for a lethal injection; 3% reported that they had written at least one prescription to be used to hasten death; and 4.7% said that they had administered at least one lethal injection. The most common reasons for the request were discomfort other than pain, loss of dignity, fear of uncontrollable symptoms, pain, and loss of meaning in life (Meier et al., 1998). Physicians were more likely to honor the requests of patients with severe pain or discomfort who had a life expectancy of less than one month and were not assessed as depressed at the time of the request (Meier et al., 2003).

Assisted suicide became lawful in Oregon in 1997, and each year approximately 0.1% of deaths in that state are by lethal prescription. One in six explicit requests for physician aid in dying are honored. Individuals who access lethal prescriptions under the law are well educated and socioeconomically secure compared with other Oregon decedents. Most patients are enrolled in home hospice when they receive the lethal prescription, suggesting that assisted suicide is not a substitute for palliative care. Physicians and hospice workers report that terminally ill individuals request assisted suicide to control the timing and manner of death and to avoid dependence on others. Maintaining independence appears to be a lifelong value for these patients. Uncontrolled pain is rarely a reason for requesting assisted suicide, though fears of worsening symptoms in the future are prominent. Depressive disorders underlie desire for hastened death in a variety of studies, but the prevalence of depression among Oregon residents who die by assisted suicide appears paradoxically low, and may represent underrecognition by clinicians (Ganzini et al., 2000, 2002; Ganzini and Dobscha, 2003). Physicians from Oregon who have received requests reported that the experience is emotionally intense, but those who agreed to participate rarely had regrets (Dobscha et al., 2004).

How should I approach euthanasia and assisted suicide in practice?

Although legal in a handful of countries and states, euthanasia and assisted suicide remain illegal and punishable by imprisonment in most jurisdictions. Physicians who believe that euthanasia and assisted suicide should be legally accepted may pursue these convictions through various legal and democratic means at their disposal: the courts and the legislature. In approaching these issues in a clinical setting, it is important to (i) thoroughly explore the reasons for the request; (ii) respect competent decisions to forgo treatment, such as discontinuing mechanical ventilation at the request

of a patient who is unable to breathe independently, which is legal; (iii) support the patient’s autonomy and attempts to maintain control in other areas of life; and (iv) provide appropriate palliative measures.

The cases

The case of Ms. I involves a competent, terminally ill patient who is imminently dying and in intractable pain. The case of Mr. J involves an apparently competent patient who is not dying but is experiencing extreme mental suffering. In both cases, the physician is confronted with a possible request to participate in euthanasia or assisted suicide. Ms. I is suffering and close to death. In consultation with her and her family, the medical team should aggressively control pain and symptoms, calling on the assistance of palliative care specialists if available. Some physicians may be concerned that this type of assertive sedation and pain management may hasten death and thus constitutes euthanasia. This approach, however, is ethically permissible as long as the goal of care is to decrease suffering, euthanasia is not the physician’s intention, and death is not the means for alleviating suffering (Williams, 2001).

In the case of Mr. J, the clinical team should explore the source of his despair and respond with psychosocial support and efforts to decrease suffering that do not end the patient’s life. Despite the absence of clinical depression, assistance from mental health experts may be beneficial.

REFEREN CES

Airedale NHS Trust v. Bland [1993] 1 All ER 821 (House of Lords, England).

Angell, M. (1997). The Supreme Court and physicianassisted suicide: the ultimate right. N Engl J Med 336: 50–3.

Annas, G. J. (1996). The promised end: constitutional aspects of physician-assisted suicide. N Engl J Med 335: 683–7.

Anon. (1996). Managing a comfortable death. Lancet 347: 1777.

Bachman, J. G., Alcser, K. H., Doukas, D. J., et al. (1996). Attitudes of Michigan physicians and the public toward legalizing physician-assisted suicide and voluntary euthanasia. N Engl J Med 334: 303–9.

Brock, D. W. (1992). Voluntary active euthanasia. Hastings Cent Rep 22: 10–22.

Brody, H. (1992). Assisted death: a compassionate response to a medical failure. N Engl J Med 327: 1384–8.

Brody, H. (1993). Causing, intending, and assisting death.

J Clin Ethics 4: 112–17.

Callaghan, D. (1992). When self-determination runs amok.

Hastings Cent Rep 22: 52–5.

Compassion in Dying v. Washington [1996] 79 F 3d 790 (9th Cir).

Dickens, B. M. (1993). Medically assisted death: Nancy B. v. Hotel–Dieu de Quebec. McGill Law J 38: 1053–70.

Dobscha, S. K., Heintz, R. T., Press, N., and Ganzini, L. (2004). Oregon physicians’ responses to requests for assisted suicide: a qualitative study. J Palliat Med 7: 450–61.

Drickamer, M. A., Lee, M. A., and Ganzini, L. (1997). Practical issues in physician-assisted suicide. Ann Intern Med 126: 146–51.

Foley, K. M. (1997). Competent care for the dying instead of physician-assisted suicide. N Engl J Med 336: 54–8.

Ganzini, L. and Dobscha, S. K. (2003). If it isn’t depression.

J Palliat Med 6: 927–30.

Ganzini, L., Nelson, H. D., Schmidt, T. A., et al. (2000). Physicians’ experiences with the Oregon Death with Dignity Act. N Engl J Med 342: 557–63.

Ganzini, L., Harvath, T. A., Jackson, A., et al. (2002). Experiences of Oregon nurses and social workers with hospice patients who requested assistance with suicide.

N Engl J Med 347: 582–8.

Genuis, S. J., Genuis, S. K., and Chang, W. C. (1994). Public attitudes toward the right to die. CMAJ 150: 701–8.

Gillon, R. (1988). Euthanasia, withholding life-prolonging treatment, and moral differences between killing and letting die. J Med Ethics 14: 115–7.

Guillod, O. and Schmidt, A. (2005). Assisted suicide under Swiss law. Eur J Health Law 12: 25–38.

Kamisar, Y. (1958). Some non-religious views against proposed ‘‘mercy killing’’ legislation. Minnesota Law Rev 42: 969–1042.

Kamisar, Y. (1995). Against assisted suicide: even a very limited form. Univ Detroit Mercy Law Rev 72: 735–69.

Kinsella, D. T. (1991). Will euthanasia kill medicine? Ann R Coll Physic Surg Canada 24: 489–92.

Kuhse, H. and Singer, P. (1988). Doctors’ practices and attitudes regarding voluntary euthanasia. Med J Aust 148: 623–7.

Lee, M. A., Nelson, H. D., Tilden, V. P., et al. (1996). Legalizing assisted suicide: views of physicians in Oregon. N Engl J Med 334: 310–15.

Meier, D. E., Emmons, C. A., Wallenstein, S., et al. (1998). A national survey of physician-assisted suicide and euthanasia in the United States. N Engl J Med 338: 1193–201.

Meier, D. E., Emmons, C. A., Litke, A., Wallenstein, S., and Morrison, R. S. (2003). Characteristics of patients requesting and receiving physician-assisted suicide.

Arch Intern Med 163: 1537–42.

Quill, T. E. (1993). The ambiguity of clinical intentions.

N Engl J Med 329: 1039–40.

Quill, T. E., Cassel, C. K., and Meier, D. E. (1992). Care of the hopelessly ill. Proposed clinical criteria for phys- ician-assisted suicide. N Engl J Med 327: 1380–4.

Quill v. Vacco [1996] 80 F 3d 716 (2nd Cir). R v. Adams (1957). Crim LR 365.

Rachels, J. (1975). Active and passive euthanasia. N Engl J Med 292: 78–80.

Rodriguez v. Attorney-General of British Columbia (1993) 107 DLR (4d) 342 (Supreme Court of Canada).

Roy, D. J. (1990). Euthanasia: where to go after taking a stand? J Palliat Care 6: 3–5.

Shapiro, R. S., Derse, A. R., Gottlieb, M., Schiedermayer, D., and Olson, M. (1994). Willingness to perform euthanasia. A survey of physician attitudes. Arch Intern Med 154: 575–84.

Singer, P. A., Choudhry, S., Armstrong, J., Meslin, E. M., and Lowy, F. H. (1995). Public opinion regarding end of life decisions: influence of prognosis, practice and process. Soc Sci Med 41: 1517–21.

Special Senate Committee on Euthanasia and Assisted Suicide (1995). Of Life or Death. Ottawa: Supply and Services Canada.

Sue Rodriguez v. British Columbia (Attorney General)

(1993) 3 SCR 519 [See Justice Cory’s dissent].

van der Maas, P. J., van der Wal, G., Haverkate, I., et al. (1996). Euthanasia, physician-assisted suicide, and other medical practices involving the end of life in the Netherlands, 1990–1995. N Engl J Med 335: 1699–705.

Veldink, J. H., Wokke, J. H., van der Wal, G., Vianney de Jong, J. M., and van den Berg, L. H. (2002). Euthanasia

and physician-assisted suicide among patients with amyotrophic lateral sclerosis in the Netherlands. N Engl J Med 346: 1638–44.

Ward, B. J. and Tate, P. A. (1994). Attitudes among NHS doctors to requests for euthanasia. BMJ 308: 1332–4.

Williams, G. (2001). The principle of double effect and terminal sedation. Med Law Rev 9: 41–53.

Wysong, P. (1996). Doctors divided on euthanasia acceptance: preference is to refer euthanasia to another doctor. Med Post 32: 1: 90.

Mrs. K, an 82-year-old woman with moderate to severe Alzheimer’s dementia, advanced heart failure, emphysema, and diabetes mellitus with neuropathy and nephropathy has just been readmitted with difficulty breathing, two days after being discharged to the care of her daughter. In the previous admission for the same problem, she was treated in the intensive care unit, narrowly avoiding intubation by the use of aggressive pulmonary toilet, antibiotics, and diuretics for possible pneumonia and congestive heart failure. Just after her second admission, the attending physician approached Mrs. K’s daughter to discuss forgoing lifesustaining treatment. ‘‘In my opinion, if your mother should have a cardiac arrest, resuscitating her would be futile,’’ said Mrs. K’s physician. The daughter reacted angrily and insisted that ‘‘everything be done,’’ because her mom is strong and can get better (as she has previously).

What is conflict in the healthcare setting at the end of life?

Conflict may be defined as disagreement between people when a decision must be made or an action taken. Healthcare providers encounter conflict in everyday practice, and one of the most difficult and distressing situations physicians face is conflict with family members over forgoing life-sustaining treatment. What should be a cooperative effort to achieve treatment goals turns into an exercise in frustration and distress.

Why is conflict in the healthcare setting at the end of life important?

In the hospital, death is routine to the caregivers, but not to patients and families. Given the emotional impact of decisions surrounding death and dying, conflicts are not surprising but are still disturbing to all parties involved and can diminish trust between doctor and patient or family. This impaired trust profoundly influences the ability of families to believe or understand the prognosis and accept physicians’ recommendations based on the patient’s goals. Physicians, meanwhile, may be angry and frustrated, distrust the family’s motives, worry about litigation, or believe that they are asked to violate their professional ethos by providing care that ‘‘does not work.’’ Although physicians and patients (or families) may disagree about the proper course of action in other settings, conflicts in the context of severe illness involve high stakes, great vulnerability, deep fears, and strongly held beliefs. The focus here is on clinician– family conflicts, and not conflicts between clinicians and patients, because when decisions about withholding or withdrawing life-sustaining treatment are contemplated, patients are often incapacitated. Furthermore, a competent patient’s wishes are justifiably given much more respect than the judgements of surrogates.

Decisions about withholding or withdrawing lifesustaining treatment can receive a great deal of publicity. While some cases, such as Terri Schiavo in the USA, reach the courts and media, most conflicts about end of life decisions do not. Even without legal or media attention, these conflicts can have serious consequences. They negatively affect the quality of care and decision making, as well as the satisfaction of both family members and healthcare providers.

Ethics

Decisions about withdrawing, withholding, or continuing life-sustaining treatment require consideration of moral as well as medical concerns. Clinicians may feel that they are violating professional norms to ‘‘do no harm’’ when they are asked to continue burdensome interventions that they consider to be of little or no benefit. Recognizing moral dimensions is an important first step, including professional obligations of compassion, respect for patients’ and families’ values and beliefs (which may differ substantially from those of the physician), competence (e.g., in prognosticating and communicating), honesty, and humility. Humility, and its antithesis arrogance, bear particular weight when families face the need to trust physicians’ prognoses and recommendations.

Law

Statutes and legal precedents from a number of jurisdictions frequently apply to end of life decision making. Many courts have addressed ‘‘right to die’’ cases permitting the withdrawing or withholding of life-sustaining treatment, although the standard of evidence required regarding what the patient would have wanted may vary. For example, the US Supreme Court decision in Cruzan v. Director, Missouri Dep’t. of Health (1990) clarified the circumstances under which a patient may refuse medical treatment or authorize another to speak for him or her, and permitted states to develop their own right to die laws. Statutes may legitimize advance directives (living wills and/or durable powers of attorney for healthcare), address the circumstances under which

a patient can refuse medical treatment, provide guidance for surrogate decision making, address physician-assisted suicide, or require healthcare organizations or doctors to inquire about advance directives. Professionals should be familiar with laws in their own country and locality, and know how to access legal advice when necessary.

Policy

Besides policies set by governments, institutions (hospitals, nursing homes, health systems) frequently include end of life issues in their policies. Some healthcare institutions have ‘‘futility’’ policies; most will have policies about withdrawing and withholding life-sustaining treatment (e.g., do not resuscitate orders) and surrogate decision making. Other relevant policies may not be formal or obvious, for instance intensive care units and emergency rooms may restrict family access to patients during certain hours or certain events (e.g., resuscitation) (Kopelman et al., 2005).

Empirical studies

End of life decision making and care have attracted an enormous amount of research, ranging from comparisons of patients’ and surrogates’ preferences to interventional studies aiming to increase advance directive use or discussions about limiting treatment (Lynn et al., 2000; Prendergast, 2001). Most of these studies portray an unfortunate reality: the wishes of patients are rarely known, poorly predicted by surrogates, unreliably followed when they are known, and patients’ symptoms remain inadequately treated (Fagerlin and Schneider, 2004; Silveira et al., 2005). The approach below integrates, when available, evidence about end of life decision making.

How should I approach conflict in the healthcare setting at the end of life in practice?

As for other problems in medicine, developing a differential diagnosis for end of life conflicts can help

80S. D. Goold, B. C. Williams, and R. Arnold

clinicians to consider carefully all of the possible explanations for the disagreement. Rather than reacting to the manifest problem (e.g., establishing code status), the first crucial step is to actively inquire what are some possible root causes of the conflict. After reflecting on root causes, addressing conflicts in end of life care should begin with a few open-ended questions. Asking the family about the patient’s past history, what other clinicians have told them about their loved one’s condition, and their choices (e.g., ‘‘Can you tell me why you are leaning toward resuscitation in the event she stops breathing?’’) allows one to identify the reasons behind decisions, as well as assess understanding. With that information, clinicians can explicitly create a differential diagnosis of the sources of conflict, which fall into three general categories: family features, healthcare provider features, and contextual (organizational and social) features. These are often present in combination and may interact.

Family features that contribute to conflicts

‘‘Family’’ refers here to a patient’s collection of intimates, who may or may not be related by blood ties. Two types of circumstance can explain the family’s role in conflict. In the first, families do not understand the medical issues. In the second, they understand the clinical situation but reach a different conclusion from healthcare providers.

Inadequate understanding of the medical situation by the family could include a completely different understanding of the prognosis. An optimistic belief that cardiopulmonary resuscitation will succeed, for instance, could reflect its 77% success rate on television (Diem et al., 1996). Consultants or nurses may inadvertently convey a different prognosis to the family than the primary physicians, so it is often useful to choose one healthcare professional to serve as the primary communicator, while other clinicians convey information through this spokesperson. Families often poorly process and imperfectly remember ‘‘bad news,’’ even when it is clearly and consistently provided. Repeating key concepts, giving written as

well as verbal information, encouraging questions (‘‘I expect you will have questions about what we discussed today. Write them down for our talk tomorrow’’) and periodically assessing understanding of the situation may improve information transfer and decrease frustration.

Denial – the inability to explicitly recognize a set of facts because of its unacceptable psychological consequences – commonly affects the ability to understand medical situations. Symptoms of denial include displacement – a focus of concern on trivial, but controllable, matters – and an inability to discuss ‘‘bad news’’ (Weissman, 2004). Mrs. K’s daughter, in the case described, asked about her mother’s oxygen level and laboratory results; conversations about the ‘‘big picture’’ were quickly turned into discussions about relatively unimportant medical processes. Effective techniques for managing denial include open-ended listening; non-defensive, neutral responses; silence; and frequent, regular opportunities for the patient or family member to communicate with a consistent healthcare provider. Reflecting and validating family members’ emotions can be especially valuable. Saying ‘‘It must be very hard for you to see your mother so ill,’’ and ‘‘You’ve been a wonderful caregiver for her for many years’’ may prompt an exchange that begins to deal with the grief, guilt, or anger that can cause denial.

Finally, healthcare language can adversely affect understanding. Problems associated with interpretation of language can be avoided by using language appropriate to the family’s educational level, by frequently assessing understanding, and by avoiding shorthand terms. Phrases like ‘‘usually,’’ ‘‘most of the time,’’ or ‘‘we cannot rule out,’’ used by physicians to convey uncertainty, may be interpreted variably (Knapp et al., 2004). Families are more likely to understand ‘‘out of 100 patients like your mother, about half will survive six months or longer’’ (Morrison, 2000) than ‘‘your mother has a 50% six month survival rate.’’ Some caution is in order, however, given the impossibility of precise prognoses for individual patients (Fox et al., 1999). Providing a range of possible outcomes can

usefully convey uncertainty, for example ‘‘Almost all patients like this will not survive to be discharged from the hospital; some die within hours or days, others might stay alive for weeks or even months with in-hospital treatments. A rare few beat all the odds.’’ Other language commonly used in discussions with families (‘‘death with dignity,’’ ‘‘everything done’’) contribute to misunderstandings. In Mrs. K’s case, her daughter may have interpreted use of the term ‘‘futile’’ to mean that the physician did not think her mother was worth treating.

Grief can contribute to an inability to make any decisions, especially decisions that may result in a loved one’s death. When Mrs. K’s daughter said ‘‘I will not be able to live without her,’’ it reflected her inability to cope with her mother’s death. Supportive, open-ended dialogue allows the family to recognize, express, and begin to work through grief. Family members’ guilt, often manifest during times of crisis, may also contribute to an unwillingness to make decisions. Guilt is recognizable when family members say, ‘‘I cannot do this,’’ or ‘‘I will not be able to live with myself.’’ Physicians may unwittingly increase feelings of guilt when they ask the family to take responsibility for medical decisions (e.g., ‘‘Do you want us to resuscitate her?’’ rather than, ‘‘What do you think your mother would want us to do?’’) (Tomlinson et al., 1990). Even if surrogates do not always make decisions that patients would make (Fagerlin and Schneider, 2004), inaccuracy in no way undermines the family’s role in decision making. Another way to treat the family’s guilt is to take responsibility for medical decisions (e.g., ‘‘You tell me about what was important to your mother, and I will recommend what we should do for her’’). The family can set positive goals and objectives of treatment (e.g., maximizing comfort) and clinicians recommend actions to achieve those goals. Clinicians should also praise family members, when appropriate, for respecting their loved one’s values and wishes at the end of life.

Occasionally, secondary gain (‘‘conflict of interest’’) may lead a family to make a decision with

which the healthcare team disagrees. Secondary gain, often suspected when conflict arises, is always present to some degree when intimates make decisions for and about each other (Goold, 2000). Identifying potential sources of secondary gain – avoiding unbearable grief, avoiding overwhelming caregiving responsibilities, or avoiding financial ruin – is nonetheless illuminating. Addressing wishes to postpone the death of a spouse because of grief and loneliness requires a very different approach, for instance, than addressing wishes to keep a spouse or parent alive to collect a pension.

Even if they understand and accept the situation, family members may make decisions with which the healthcare team disagrees. Clinicians’ values and those of patients or families differ. Individuals may have vastly different ideas about what constitutes a reasonable chance worth pursuing, a good quality of life, or a ‘‘good death.’’ If Mrs. K’s daughter, deciding according to her best understanding of her mother’s wishes, chooses resuscitation because it might prolong her mother’s life for a few days, weeks, or months even though the chance of survival to discharge is very small, this decision probably reflects a difference in values and should be respected. Empirical data suggest that these conflicts occur infrequently; with good communication, doctors and families usually come to mutually agreeable moral decisions.

Healthcare provider features

Clinicians, like patients, may be uncomfortable with prognostic uncertainty (Spikes and Holland, 1975; Kahneman et al., 1982; Novack et al., 1997; Christakis and Lamont, 2000; Meier et al., 2001), which may lead them to approach limiting treatment decisions in overly hesitant or overly confident ways. Statements like ‘‘She won’t leave the hospital alive’’ or ‘‘She has less than six months to live’’ fail to take into account the near-universal uncertainty in prognosis (Christakis and Lamont, 2000), and can make families suspicious they are not being told the whole story. Likewise, communicating information or recommendations too

82S. D. Goold, B. C. Williams, and R. Arnold

vaguely can lead to confusion or false hope. Clinicians who have known the patient for a long time might provide overly optimistic prognoses.

Like patients and families, healthcare providers are often uncomfortable discussing death; anxiety about one’s own mortality may lead clinicians to avoid frank discussions about death or to provide false reassurance that ‘‘everything is OK.’’ Clinicians also face the troubling thought of a medical ‘‘failure’’ (Spikes and Holland, 1975). Healthcare providers tend to underestimate the quality of life of chronically ill patients, especially for demented patients, and are more likely than patients or families to think that such patients would choose to forgo life-sustaining treatment. Other clinician attitudes that influence conflict include beliefs about the sanctity of life, the proper role of family members, difficulty with radically different values, or insecurity about one’s competence or skill.

Insight into one’s own limitations and beliefs helps clinicians to understand feelings of anger and frustration with certain families and then to discuss with the family areas of disagreement (Novack et al., 1997; Meier et al., 2001).

Similarly, knowledge or skill deficits can catalyze clinician–family conflicts. Clinicians may be unaware of the prognosis or treatment options and misinform the family, although now numerous resources provide information on prognostic indicators for patients with a variety of clinical conditions (Gage et al., 2000). They may not understand ethical, legal, or hospital policies surrounding end of life care. Mistaken beliefs regarding the legality of withdrawing ventilators or artificial nutrition, for example, may lead a provider to refuse to accede to the family’s desire. A lack of training in palliative care and symptom management can make interventions more burdensome for the patient, and hence lead clinicians to perceive they are inflicting suffering. Finally, healthcare providers may be illtrained in interpersonal communication regarding end of life decisions, leading to misunderstandings, confusion, and frustrations (Tulsky et al., 1995, 1998). Fortunately, skills training is now more widely available.

Healthcare professionals, like patients and families, can be overworked, fatigued, frustrated, stressed, and otherwise beset by competing concerns. Physicians in training have heavy workloads and may be poorly motivated to spend additional time with patients or care for more of them and hence especially intent on making decisions quickly. The intern caring for Mrs. K may feel that her scarcest resource, time, is ‘‘wasted’’ on a demented, terminally ill woman. The culture of the hospital, with its prioritization of emergency, life and death decisions, high technology, and speedy discharges, as well as (in some systems) poor reimbursement for conversations with families, contributes to an emphasis on ‘‘high-tech’’ interventions and the avoidance of time-consuming family conferences. Insight into one’s emotional status may help, although larger cultural changes in medicine will probably be needed (Scott et al., 1995; Field and Cassel, 1997; Mildred and Solomon, 2000).

Social and organizational features

Both the immediate and the general context in which clinician–family communication occurs can influence conflict. Conversations about end of life decisions that are unexpected (to the family), unannounced, or unplanned are more likely to result in conflict than those preceded by preliminary communication between clinicians and family, that occur at a preplanned time and location, and that have agreed-upon participants. As no competent surgeon would begin an operation without a plan for the procedure, clinicians should enter family discussions prepared with information about prognoses and with prepared methods to communicate information and ask and answer questions. The actual participants are important. Trust, which often accompanies long-standing doctor–patient relationships, can be invaluable for effective, solution-oriented communication. With the increasing use of hospitalists, seeking the input and participation of continuity clinicians may help in communication.

System-wide social and organizational factors also contribute to conflicts. Hospitals and health systems are worried about finances; doctors are under pressure to constrain the use of limited resources, and there is a pervasive social feeling that too much money is spent on medical care. Consequently, when a patient’s prognosis seems to be hopeless, clinicians may feel, on the one hand, that life-prolonging treatment should be stopped quickly. On the other hand, some incentives such as reporting mortality rates for surgeons may encourage physician over treatment.

Patients and families face economic pressures of their own from serious illness; in the USA, medical expenses are the most common cause of personal bankruptcy. Families may prolong inpatient treatment because it costs them less than caring for a patient at home, or end things quickly if they bear substantial financial costs of care. Sadly, economic circumstances predict greater suffering at the end of life, and the availability of hospice care can depend on where a patient lives, their particular diagnosis, and whether or not they are insured (Silveira et al., 2003, 2005).

Hospital policies may also promote conflicts in end of life decisions. When intensive care and other units restrict visiting hours, this minimizes contact between families and patients and may impair communication between them, and keeps families from seeing what their loved one is going through (Rosenczweig, 1998; Kopelman et al., 2005). Similarly, requiring physicians to sign orders limiting life-support treatments may lead to unwanted resuscitation, particularly in settings (e.g., nursing facilities) staffed primarily by other clinicians. Organizations as well as doctors have legal fears regarding end of life decisions. In most of the major court cases in the USA since the mid 1980s, organizations refused to accede to family wishes, leading to legal action.

By considering this list of potential sources of conflict, clinicians can more readily and accurately identify the causes of difficult interactions with families of desperately ill patients around decisions to limit treatment. Improving the quality of end of

life care requires development and research in interventions designed to identify and decrease these sources of conflicts. Training may help to improve physicians’ capacity to elicit and identify psychological and social factors at play in conflicts at the end of life (Smith et al., 1998) and improve their ability to give bad news, deal with emotions, and negotiate treatment goals. Hospitals and healthcare organizations should also experiment with structural changes, such as changing visiting time or increasing support for family meetings, to minimize conflicts and facilitate acceptable and relatively efficient solutions. It is hoped that by more accurately identifying the ‘‘diagnosis,’’ the effective ‘‘treatment’’ (empathic end of life care) and ‘‘prevention’’ (early clinician–patient discussions and institutional change) will follow.

The case

Mrs. K was intubated and transferred to the intensive care unit when she experienced respiratory distress. Her daughter received counseling from clergy. She consistently expressed a request that her mother’s treatment not exclude the goal of extending life. Several days after transfer to the intensive care unit, Mrs. K died.

REFEREN CES

Christakis, N. A. and Lamont, E. B. (2000). Extent and determinants of error in doctors’ prognoses in terminally ill patients: prospective cohort study. BMJ 320: 469–73.

Cruzan v. Director, Missouri Dep’t. of Health [1990] 110 S. Ct. 2841.

Diem, S. J., Lantos, J. D., and Tulsky, J. A. (1996). Cardiopulmonary resuscitation on television. Miracles and misinformation. N Engl J Med 334: 1578–82.

Fagerlin, A. and Schneider, C. E. (2004). Enough: the failure of the living will. Hastings Cent Rep 34: 30–42.

Field, M. J. and Cassel, C. K., for the Committee on Care at the End of Life of the Institute of Medicine (1997).

Approaching Death: Improving Care at the End of Life. Washington, DC: Institute of Medicine.